TWI644563B - Method of generating quantized block - Google Patents

Abstract

One method is to use the most probable mode group to derive the in-frame prediction mode of the current prediction unit. The MPM group includes three in-frame prediction modes determined by the left-in-frame prediction mode of the current prediction unit and the in-frame prediction mode in the upper image. Significant markers, coefficient symbols, and coefficient levels are inversely scanned according to the selected inverse scan mode to produce a quantized block. When the in-frame prediction mode on the left and the in-frame prediction mode on the top are the same as each other and the in-frame prediction mode on the left is one of two non-directional in-frame prediction modes, the MPM group includes two non-directional in-frame prediction modes. With vertical mode. When only one of the in-frame prediction modes on the left and the in-frame prediction modes on the top are available and one of the multiple directional in-frame prediction modes, the MPM group includes two non-directional in-frame prediction modes and the available in-frame prediction modes. Forecasting mode.

Description

Method for generating quantized blocks

The present invention relates to a method and a device for decoding an image, and more specifically, the present invention relates to a method and a device for generating a quantized block by adaptively determining an inverse scanning mode based on an intra-frame prediction mode and a size of a transform unit.

In H.264 / MPEG-4 AVC, a picture is divided into multiple macroblocks to encode an image, and inter-frame prediction or intra-frame prediction is used to generate a prediction block to encode the corresponding macroblock. The difference between the initial block and the prediction block is transformed to generate a transformed block, and the transformed block is quantized using a quantization parameter and one of a plurality of predetermined quantization matrices. The quantization coefficients of the quantized block are scanned through a predetermined scan type and then entropy-coded. The quantization parameter is adjusted for each macroblock, and it is encoded using the previous quantization parameter.

At the same time, technologies using various sizes of coding units and transformation units are introduced to improve coding efficiency. Techniques to increase the number of prediction modes within the frame were also introduced to produce prediction blocks that are more similar to the initial block.

However, when scanning a large transform block, various sizes of the coding unit and the transform unit cause the coding bits of the residual block to increase. Moreover, increasing the number of prediction modes within the frame requires a more efficient scanning method to reduce the coding bits of the residual block.

The present invention relates to a method of deriving an intra-frame prediction mode of a prediction unit; selecting an inverse scan mode of a current transform unit based on the in-frame prediction mode and the size of a transform unit; and inversely scanning a marked mark according to the selected inverse scan mode, Coefficient signs and coefficient levels to produce quantized blocks.

An aspect of the present invention provides a method for generating a quantized block, which includes: using a most probable mode group to derive an in-frame prediction mode of a current prediction unit, the MPM group including the left-hand frame according to the current prediction unit The three in-frame prediction modes determined by the prediction mode and the in-frame prediction mode in the figure above; selecting an inverse scanning mode based on the in-frame prediction mode and the size of a transform unit; and inverse-scanning the marked mark based on the selected inverse scanning mode, Coefficient signs and coefficient levels to generate quantized blocks, where the inverse scan mode is one of diagonal scan, vertical scan, and horizontal scan. When the in-frame prediction mode on the left is the same as the in-frame prediction mode on the top and the in-frame prediction mode on the left is one of the two non-directional in-frame prediction modes, the MPM group includes two types of non-directional in-frame prediction. Mode and a vertical mode. When only one of the in-frame prediction modes on the left and the in-frame prediction modes on the above are available and are one of the multiple directional in-frame prediction modes, the MPM group contains two non-directional in-frame prediction modes and one available. In-frame prediction mode.

The in-frame prediction mode of the prediction unit is derived according to the method of the present invention; the inverse scanning mode of the current transformation unit is selected between diagonal scanning, vertical scanning, and horizontal scanning based on the in-frame prediction mode and the size of the transformation unit; and through Significant markers, coefficient signs, and coefficient levels are inversely scanned according to the selected inverse scan mode to produce a quantized block. If the transform unit is larger than a predetermined size, multiple subsets are generated and inversely scanned. Therefore, the scan mode is determined by determining the scan unit based on the transform unit size and the intra-frame prediction mode, and the amount of coding bits of the residual block is reduced by applying the scan mode to each subset.

Hereinafter, different embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, the present invention is not limited to the exemplary embodiments disclosed below, but may be implemented in various ways. Therefore, many other modifications and variations of the present invention are possible, and it is understood that within the scope of the disclosed concept, the present invention can be practiced in a manner different from that specifically described.

FIG. 1 is a block diagram of an image encoding device 100 according to the present invention.

Please refer to FIG. 1. The image encoding device 100 according to the present invention includes a picture division unit 101, a transformation unit 103, a quantization unit 104, a scanning unit 105, an entropy encoding unit 106, an inverse quantization unit 107, An inverse transform unit 108, a post-processing unit 110, a picture storage unit 111, an intra-frame prediction unit 112, an inter-frame prediction unit 113, a subtracter 102, and an adder 109.

The picture division unit 101 divides a picture or slice into a plurality of maximum coding units (LCUs), and divides each LCU into one or more coding units. The picture division unit 101 determines a prediction mode and a prediction unit size and a transformation unit size of each coding unit.

The LCU contains one or more coding units. The LCU has a recursive quad-tree structure to specify the partition structure of the LCU. Information specifying the maximum size and minimum size of the coding unit is included in the sequence parameter set. The split structure is specified by one or more split coding unit flags (split_cu_flags). The size of the coding unit is 2N × 2N.

A coding unit contains one or more prediction units. In the in-frame prediction, the size of the prediction unit is 2N × 2N or N × N. In inter-frame prediction, the sizes of prediction units are 2N × 2N, 2N × N, N × 2N, and N × N. When the prediction unit is asymmetrically partitioned in the inter-frame prediction, the size of the prediction unit may also be one of hN × 2N, (2-h) N × 2N, 2N × hN, and 2N × (2-h) N. The value of H is 1/2.

A coding unit contains one or more transformation units. The transformation unit has a recursive quad-tree structure to specify the segmentation structure. The split structure is specified by one or more split transform unit flags (split_tu_flags). Information specifying the maximum and minimum size of the transformation unit is included in the sequence parameter set.

The in-frame prediction unit 112 determines the in-frame prediction mode of the current prediction unit and uses the in-frame prediction mode to generate one or more prediction blocks. The prediction block has the same size as the transform unit.

FIG. 2 is a schematic diagram of an in-frame prediction mode according to the present invention. As shown in FIG. 2, the number of prediction modes in the frame is 35. The DC mode and the plane mode are non-directional in-frame prediction modes, and the other are directional in-frame prediction modes.

The inter-frame prediction unit 113 uses one or more reference pictures stored in the picture storage unit 111 to determine motion information of the current prediction unit and generates a prediction block of the prediction unit. The motion information includes one or more reference picture indexes and one or more motion vectors.

The transform unit 103 transforms the residual signal using the initial block and the prediction block to generate a transform block. The residual signal is transformed by a transform unit. The transform type is determined by the prediction mode and the size of the transform unit. The transform type is a DCT-based integer transform or a DST-based integer transform.

The quantization unit 104 determines a quantization parameter for quantizing the transform block. The quantization parameter is the quantization step size. A quantization parameter is determined for each quantization unit. The size of the quantization unit is one of the allowable sizes of the coding unit. If the size of the coding unit is equal to or larger than the minimum size of the quantization unit, the coding unit is set as a quantization unit. The quantization unit may include multiple coding units. The minimum size of the quantization unit is determined for each picture, and a parameter is used to specify the minimum size of the quantization unit. Include parameters in the picture parameter set.

The quantization unit 104 generates a quantization parameter predictor and generates a differential quantization parameter by subtracting the quantization parameter predictor from the quantization parameter. The differential quantization parameters are encoded and sent to a decoder. If there is no residual signal to be transmitted within the coding unit, the differential quantization parameter of the coding unit may not be transmitted.

A quantization parameter predictor is generated using the quantization parameters of neighboring coding units and the quantization parameters of a previous coding unit as follows.

The left quantization parameter, the upper quantization parameter, and the pre-quantization parameter are sequentially searched in the following order. When there are two or more quantization parameters, the average of the first two available quantization parameters retrieved in that order is set as the quantization parameter predictor, and when there is only one quantization parameter, the available quantization parameters are set as the quantization parameters Predictor. That is, if there are left and up quantization parameters, the average of the left and up quantization parameters is set as a quantization parameter predictor. If there is only one of the left and upper quantization parameters, the average of the available quantization parameter and the previous quantization parameter is set as the quantization parameter predictor. If neither the left nor upper quantization parameters are available, the previous quantization parameter is set as the quantization parameter predictor. Round the average.

The differential quantization parameter is converted into an absolute value of the differential quantization parameter and a symbol mark representing a symbol of the differential quantization parameter. Binary the absolute value of the differential quantization parameter to a truncated unary code. Then, the absolute value and the sign label are arithmetically coded. If the absolute value is zero, there is no symbol tag.

The quantization unit 104 quantizes the transform block by using a quantization matrix and a quantization parameter. The inverse quantization unit 107 and the scanning unit 105 are provided with a quantization block.

The scanning unit 105 determines a scanning mode and applies a scanning mode to the vectorized block.

In the in-frame prediction, the distribution of the quantized transform coefficients changes according to the in-frame prediction mode and the size of the transform unit. Then, the scanning mode is determined by the intra prediction mode and the size of the transform unit. The size of the transform unit, the size of the transform block, and the size of the quantization block are the same.

FIG. 3 is a schematic diagram of a scanning mode according to the present invention. FIG. 4 is a schematic diagram of a diagonal scan according to the present invention. As shown in FIG. 3, the first scanning mode is a zigzag scanning, the second scanning mode is a horizontal scanning, and the third scanning mode is a vertical scanning.

When CAVLC (Context Adaptive Variable Length Coding) is used for entropy coding, a scanning mode is selected between zigzag scanning, horizontal scanning, and vertical scanning. However, when using CABAC (Context Adaptive Binary Arithmetic Coding) for entropy coding, a scanning mode is selected between diagonal scanning, horizontal scanning and vertical scanning, and the selected scanning mode is applied to the Notable marks, coefficient symbols, and coefficient levels. The significant flag indicates whether the corresponding quantized transform coefficient is zero. The coefficient symbol indicates the sign of the non-zero quantized transform coefficient, and the coefficient level indicates the absolute value of the non-zero quantized transform coefficient.

"Figure 5" is determined by the in-frame prediction mode and transform unit size according to the present invention. Exemplary scan mode. When CABAC is used for entropy coding, the scanning mode is determined as follows.

When the size of the transform unit is 4 × 4, horizontal scanning is applied to the prediction mode of the adjacent frame in the vertical mode (mode 1) and the first number of vertical modes, and the first number of the horizontal mode (mode 2) and the horizontal mode. The vertical prediction mode is applied to the adjacent in-frame prediction modes of, and a diagonal scan is applied to all other in-frame prediction modes. That is, if it is assumed that the allowable intra-frame prediction mode for 4 × 4 is modes 0 to 17, horizontal scanning is applied for modes 5, 6 and modes allowed between modes 5 and 6, and for modes 8, Mode 9 and modes allowed between modes 8 and 9 apply vertical scanning. If the allowed intra-frame prediction mode for 4 × 4 is modes 0 to 34, the scanning mode applied is the same as the following 8 × 8 transform unit.

When the size of the transform unit is 8 × 8, horizontal scanning is applied to the in-frame prediction mode of the second number of adjacent modes of the vertical mode (mode 1) and the vertical mode, and the second number of the horizontal mode (mode 2) and the horizontal mode. The vertical prediction mode is applied to the adjacent in-frame prediction modes of, and a diagonal scan is applied to all other in-frame prediction modes. That is, horizontal scanning is applied for modes 5, 6 and modes allowed between modes 5 and 6, and vertical scanning is allowed for modes 8, 9 and modes allowed between modes 8 and 9, for all other frames The prediction mode applies a diagonal scan. The modes allowed between Mode 5 and Mode 6 are Modes 21, 12, 22, 1, 23, 13, and 24. The modes allowed between Mode 8 and Mode 9 are Modes 29, 16, 30, 2, 31, 17, 32, and 9.

In the inter-frame prediction, a predetermined scanning mode is used regardless of the size of the transform unit. When CABAC is used for entropy coding, the predetermined scanning mode is diagonal scanning.

When the size of the transform unit is larger than the second size, the quantization block is divided into a main subset and a plurality of remaining subsets, and the determined scanning mode is applied to each subset. According to the determined scanning mode, the salient mark, coefficient symbol and coefficient level of each subset are scanned separately. Divide the quantized transform coefficients into salient marks, coefficient signs, and coefficient levels.

The main subset contains DC coefficients, and the remaining subset covers areas other than the area covered by the main subset. The second size is 4 × 4. The size of the subset can be 4 × 4 blocks or non-square blocks determined by the scan mode. Non-square blocks contain 16 transform coefficients. For example, the size of the subset is 8 × 2 for horizontal scanning, 2 × 8 for vertical scanning and 4 × 4 for diagonal scanning.

The scan mode used to scan the subsets is the same as the scan mode used to scan the quantized transform coefficients of each subset. The quantized transform coefficients of each subset are scanned in opposite directions. The subset is also scanned in the reverse direction.

The last non-zero coefficient position is encoded and sent to the decoder. The last non-zero coefficient position specifies the position of the last non-zero quantized transform coefficient in the transform unit. The last non-zero coefficient position is used to determine the number of subsets sent in the decoder. A non-zero subset flag is set for each subset except the main subset and the last subset. The last subset covers the last non-zero coefficient. The non-zero subset flag indicates whether the subset contains non-zero coefficients.

The inverse quantization unit 107 performs inverse quantization on the transform coefficients quantized by the quantization block.

The inverse transform unit 108 inversely transforms the inverse quantized block to generate a residual signal in the spatial domain.

The adder 109 generates a reconstructed block by adding the residual block and the prediction block.

The post-processing unit 110 executes a deblocking filtering process to remove block artifacts generated in the reconstructed picture.

The picture storage unit 111 receives a post-processed image from the post-processing unit 110 and stores the image in the picture unit. The picture can be a frame or a field.

The entropy encoding unit 106 performs entropy encoding on the one-dimensional coefficient information received from the scanning unit 105, the in-frame prediction information received from the in-frame prediction unit 112, the motion information received from the inter-frame prediction unit 113, and the like.

FIG. 6 is a block diagram of an image decoding device 200 according to the present invention.

The image decoding device 200 according to the present invention includes an entropy decoding unit 201, an inverse scanning unit 202, an inverse quantization unit 203, an inverse transform unit 204, an adder 205, a post-processing unit 206, and a picture storage unit 207 An intra-frame prediction unit 208 and an inter-frame prediction unit 209.

Entropy decoding unit 201 extracts in-frame prediction information and inter-frame prediction information from the received bit stream Forecast information and one-dimensional coefficient information. The entropy decoding unit 201 sends inter-frame prediction information to the inter-frame prediction unit 209, sends the intra-frame prediction information to the intra-frame prediction unit 208, and sends coefficient information to the inverse scanning unit 202.

The inverse scanning unit 202 generates a quantized block using an inverse scanning mode. When CABAC is used for entropy coding, the scanning mode is determined as follows.

Select the inverse scan mode between diagonal scan, vertical scan, and horizontal scan.

In the in-frame prediction, the inverse scanning mode is determined by the in-frame prediction mode and the size of the transform unit. Select the inverse scan mode between diagonal scan, vertical scan, and horizontal scan. The selected inverse scanning mode is applied to the salient mark, the coefficient sign, and the coefficient level to generate a quantized block.

When the size of the transformation unit is equal to or smaller than the first size, a horizontal scan is selected for a predetermined number of adjacent in-frame prediction modes of the vertical mode and a vertical mode, and a predetermined number of adjacent in-frame prediction modes are selected for the horizontal mode and the horizontal mode Vertical scan, select diagonal scan for other in-frame prediction modes. When the size of the transform unit is larger than the first size, a diagonal scan is used. When the size of the transform unit is larger than the first size, a diagonal scan is selected for all in-frame prediction modes. The first size is 8 × 8.

When the size of the transformation unit is 4 × 4, horizontal scanning is applied for the vertical mode (mode 1) and the first number of adjacent in-frame prediction modes with the closest direction to the vertical mode, and the horizontal mode (mode 2) and the The horizontal mode applies the vertical mode to the first number of adjacent intra-frame prediction modes with the nearest direction, and applies diagonal scanning to all other intra-frame prediction modes. That is, if it is assumed that the allowable intra-frame prediction mode for 4 × 4 is modes 0 to 17, horizontal scanning is applied for modes 5, 6 and modes allowed between modes 5 and 6, and for modes 8, Mode 9 and modes allowed between modes 8 and 9 apply vertical scanning. If the allowed intra-frame prediction mode for 4 × 4 is modes 0 to 34, the scanning mode applied is the same as the following 8 × 8 transform unit.

When the size of the transform unit is 8 × 8, horizontal scanning is applied to the vertical mode (mode 1) and the second number of adjacent in-frame prediction modes with the closest direction to the vertical mode, and the horizontal mode (mode 2) and to Horizontal mode applies the vertical mode to the second number of adjacent intra-frame prediction modes with the nearest direction, and applies diagonal scanning to all other intra-frame prediction modes. That is, mode 5 and mode 6 And horizontal scanning is allowed between modes 5 and 6, horizontal scanning is allowed for modes 8, 9 and 8 and 9, and diagonal scanning is applied for all other in-frame prediction modes. The modes allowed between Mode 5 and Mode 6 are Modes 21, 12, 22, 1, 23, 13, and 24. The modes allowed between mode 8 and mode 9 are modes 29, 16, 30, 2, 31, 17, 32, and 9.

In interframe prediction, diagonal scanning is used.

When the size of the transformation unit is larger than the second size, the determined inverse scanning mode is used to inversely scan the significant tags, coefficient symbols, and coefficient levels in the units of the subset to generate a subset, and inversely scan the subset to generate a quantized block. The second size is 4 × 4. The size of the subset can be 4 × 4 blocks or non-square blocks determined by the scan mode. Non-square blocks contain 16 transform coefficients. For example, the size of the subset is 8 × 2 for horizontal scanning, 2 × 8 for vertical scanning, and 4 × 4 for diagonal scanning.

The inverse scan mode used to generate each subset is the same as the inverse scan mode used to generate quantized blocks. Scan the salient markers, coefficient signs, and coefficient levels in the opposite direction. The subset is also inversely scanned in the reverse direction.

The last non-zero coefficient position and a non-zero subset flag are received from the decoder. The number of coding subsets is determined according to the last non-zero coefficient position and the inverse scanning mode. Use a non-zero subset flag to select the subset to be generated. The inverse scan mode is used to generate the main and final subsets.

The inverse quantization unit 203 receives the differential quantization parameter from the entropy decoding unit 201 and generates a quantization parameter predictor. A quantization parameter predictor is generated by the same operation of the quantization unit 104 of FIG. 1. Then, the inverse quantization unit 203 adds the differential quantization parameter and the quantization parameter predictor to generate a quantization parameter of the current coding unit. If the size of the current coding unit is equal to or larger than the minimum size of the quantization unit and the differential quantization parameter for the current coding unit is not received from the encoder, the differential quantization parameter is set to 0.

A quantization parameter is generated for each quantization unit. If the size of the coding unit is equal to or greater than the minimum size of the quantization unit, a quantization parameter is generated for the coding unit. If the quantization unit includes multiple coding units, a quantization parameter is generated for a first coding unit that includes one or more non-zero coefficients in a decoding order. The coding unit after the first coding unit within the quantization unit has the same quantization parameter as the first coding unit.

Use only one parameter included in the picture parameter set and the size of the maximum coding unit to derive the minimum size of the quantization unit for each picture.

The differential quantization parameters are restored for each quantization unit. Arithmetically decode the encoded differential quantization parameter to generate an absolute value of the differential quantization parameter and a symbol label representing a symbol of the differential quantization parameter. The absolute value of the differential quantization parameter is a truncated binary string of a unary code. Then, the differential quantization parameters are restored by using the absolute value and the sign mark. If the absolute value is zero, there is no symbol tag.

The inverse quantization unit 203 performs inverse quantization on the quantized block.

The inverse transform unit 204 inversely transforms the inversely quantized block to restore the residual block. The inverse transform type is adaptively determined according to the prediction mode and the size of the transform unit. The inverse transform type is a DCT-based integer transform or a DST-based integer transform.

The in-frame prediction unit 208 restores the in-frame prediction mode of the current prediction unit by using the received in-frame prediction information, and generates a prediction block according to the restored in-frame prediction mode.

The inter-frame prediction unit 209 uses the received inter-frame prediction information to recover the motion information of the current prediction unit, and uses the motion information to generate a prediction block.

The post-processing unit 206 operates similarly to the post-processing unit 110 of FIG. 1.

The picture storage unit 207 receives the post-processed image from the post-processing unit 206 and stores the image in the picture unit. The picture can be a frame or a field.

The adder 205 adds the restored residual block to the prediction block to generate a reconstructed block.

FIG. 7 is a flowchart of a method for generating a prediction block according to the present invention.

Entropy decoding is performed on the prediction information in the frame of the current prediction unit (S110).

The prediction information in the frame includes a mode group indicator and a prediction mode index. The mode group indicator is a flag indicating whether the prediction mode in the frame of the current prediction unit belongs to the most likely mode group (MPM group). If the flag is 1, the in-frame prediction unit of the current prediction unit belongs to the MPM group. If the flag is 0, the prediction unit in the frame of the current prediction unit belongs to the residual mode group. The residual mode group includes all intra-frame prediction modes except the intra-frame prediction modes belonging to the MPM group. The prediction mode index specifies the intra-frame prediction mode of the current prediction unit within the group specified by the mode group indicator.

The MPM group is constructed using the intra-frame prediction mode of the neighboring prediction unit (S120).

The intra-frame prediction mode of the MPM group is adaptively determined from the intra-frame prediction mode on the left and the intra-frame prediction mode on the upper image. The in-frame prediction mode on the left is the in-frame prediction mode of the left adjacent prediction unit, and the in-frame prediction mode on the upper is the in-frame prediction mode of the upper adjacent prediction unit. The MPM group consists of three intra-frame prediction modes.

If there is no left or upper neighboring prediction unit, the in-frame prediction mode of the left or upper neighboring unit is set to unavailable. For example, if the current prediction unit is located at the left or upper boundary of the picture, there are no left or upper neighboring prediction units. If the left or upper neighboring unit is located in another slice or other tile, then the in-frame prediction mode of the left or upper neighboring unit is set to unavailable. If the left or upper neighboring units are mutually coded, the intra-frame prediction mode of the left or upper neighboring units is set to unavailable. If the upper neighboring unit is located within another LCU, set the intra-frame prediction mode of the left or upper neighboring unit to unavailable.

When both the left and right in-frame prediction modes are available and different from each other, include the left and right in-frame prediction modes in the MPM group and add an extra in-frame prediction mode Add to MPM group. The index 0 is assigned to one intra-frame prediction mode with a small mode number, and the index 1 is assigned to another. Or, the index 0 is assigned to the in-frame prediction mode in the left figure, and the index 1 is assigned to the in-frame prediction mode in the upper figure. The increased in-frame prediction mode is determined from the left and upper in-frame prediction modes as follows.

If one of the prediction modes in the frame on the left and the top is a non-directional mode and the other is a directional mode, the other non-directional mode is added to the MPM group. For example, if one of the in-frame prediction modes on the left and above is a DC mode, a planar mode is added to the MPM group. If one of the prediction modes in the frame on the left and the above is a flat mode, the DC mode is added to the MPM group. If both the left and top frame prediction modes are non-directional, add the vertical mode to the MPM group. If the in-frame prediction modes on the left and above are both directional modes, add DC mode or planar mode to the MPM group.

When there is only one of the in-frame prediction mode on the left and the in-frame prediction mode on the above, the available in-frame prediction modes are included in the MPM group, and the other two in-frame prediction modes are added to the MPM group. The two added intra-frame prediction modes are determined through the available intra-frame prediction modes as follows.

If the available intra-frame prediction mode is a non-directional mode, other non-directional modes and vertical modes are added to the MPM group. For example, if the available intra-frame prediction mode is the DC mode, the planar mode and the vertical mode are added to the MPM group. If the available intra-frame prediction mode is a flat mode, then the DC mode and the vertical mode are added to the MPM group. If the available intra-frame prediction mode is a directional mode, two non-directional modes (DC mode and planar mode) are added to the MPM group.

When both the in-frame prediction mode on the left and the in-frame prediction mode on the above are available and are the same as each other, the available in-frame prediction modes are included in the MPM group and two additional in-frame prediction modes are added to the MPM group. The two added intra-frame prediction modes are determined through the available intra-frame prediction modes as follows.

If the available intra-frame prediction mode is a directional mode, two adjacent directional modes are added to the MPM group. For example, if the available intra-frame prediction mode is mode 23, the left adjacent mode (mode 1) and the right adjacent mode (mode 13) are added to the MPM group. If the available intra-frame prediction mode is mode 30, add two adjacent modes (mode 2 and mode 16) to the MPM group. If the available intra-frame prediction mode is a non-directional mode, other non-directional modes and vertical modes are added to the MPM group. For example, if the available intra-frame prediction mode is the DC mode, the planar mode and the vertical mode are added to the MPM group.

When neither the in-frame prediction mode in the left picture nor the in-frame prediction mode in the upper picture is available, three additional in-frame prediction modes are added to the MPM group. The three intra-frame prediction modes are DC mode, planar mode, and vertical mode. The indices 0, 1, and 2 are assigned to the three intra-frame prediction modes in the order of the DC mode, the planar mode, and the vertical mode or the order of the planar mode, the DC mode, and the vertical mode.

It is determined whether the mode group indicator indicates an MPM group (S130).

If the mode group indicator indicates the MPM group, the in-frame prediction of the MPM group specified by the prediction mode index is set as the in-frame prediction mode of the current prediction unit (S140).

If the mode group indicator does not indicate the MPM group, the intra-frame prediction is derived by comparing the prediction mode index of the MPM group and the intra-frame prediction mode as in the following ordered steps (S150).

1) Among the three in-frame prediction modes of the MPM group, the in-frame prediction mode with the lowest mode number is set as the first candidate, the in-frame prediction mode in the middle of the mode number is set as the second candidate, and the mode number is The highest intra-frame prediction mode is set as the third candidate.

2) Compare the prediction mode index with the first candidate. If the prediction mode index is equal to or greater than the first candidate of the MPM group, the value of the prediction mode index is increased by one. Otherwise, the value of the prediction mode index is maintained.

3) Compare the prediction mode index with the second candidate. If the prediction mode index is equal to or greater than the second candidate of the MPM group, the value of the prediction mode index is increased by one. Otherwise, the value of the prediction mode index is maintained.

4) Compare the prediction mode index with the third candidate. If the prediction mode index is equal to or greater than the third candidate of the MPM group, the value of the prediction mode index is increased by one. Otherwise, the value of the prediction mode index is maintained.

5) Set the value of the last prediction mode index to the mode number of the prediction mode within the frame of the current prediction unit.

The size of the prediction block is determined based on the transform size indicator of the designated transform unit size (S160). The transform size indicator can be a split_transform_flag that specifies the size of the transform unit.

If the size of the transformation unit is equal to the size of the current prediction unit, a prediction block is generated through the following steps S170 to S190.

If the size of the transform unit is smaller than the size of the current prediction unit, the prediction block of the first sub-block of the current prediction unit is generated through steps S170 to S190, and the residual block and the first of the first current sub-block are generated by adding the prediction block and the residual block. The reconstructed block of the child block. Then, a reconstructed block of the next sub-block in the decoding order is generated. Use the same in-frame prediction mode for all sub-blocks. A sub-block has the size of a transformation unit.

It is determined whether all reference pixels of the current block are available, and if one or more reference pixels are not available, a reference pixel is generated (S170). The current block is the current prediction unit or the current sub-block. The size of the current block is the size of the transformation unit.

The reference pixels are adaptively filtered based on the intra-frame prediction mode and the size of the current block (S180). The size of the current block is the size of the transformation unit.

In DC mode, vertical mode, and level mode, the reference pixels are not filtered. In directional modes other than vertical and horizontal modes, the reference pixels are adjusted according to the size of the current block.

If the current size is 4 × 4, the reference pixels are not filtered in all in-frame prediction modes. For the sizes of 8 × 8, 16 × 16, and 32 × 32, as the current block size becomes larger, the number of in-frame prediction modes for filtering reference pixels increases. For example, the reference pixels are not filtered in the vertical mode and a predetermined number of adjacent intra-frame prediction modes in the vertical mode. Reference pixels are also not filtered in the horizontal mode and the predetermined number of adjacent intra-frame prediction modes of the horizontal mode. The predetermined number increases as the current block size increases.

The reference pixels of the current prediction unit and the intra-frame prediction mode are used to generate a prediction block of the current block (S190).

In the vertical mode, the value of a vertical reference pixel is copied to generate a predicted pixel. The corner reference pixels and the left adjacent reference pixels are used to filter the predicted pixels adjacent to the left reference pixel.

In the horizontal mode, the value of a horizontal reference pixel is copied to generate a predicted pixel. The corner reference pixels and the upper adjacent reference pixels are used to filter the predicted pixels adjacent to the upper reference pixels.

Although the invention has been shown and described with reference to certain exemplary embodiments thereof, those skilled in the art will understand that changes can be made in various forms and details without departing from the spirit and scope of the invention as defined by the scope of the patent application .

100‧‧‧Image coding equipment

101‧‧‧Screen Segmentation Unit

102‧‧‧Subtractor

103‧‧‧ Transformation unit

104‧‧‧Quantization unit

105‧‧‧scanning unit

106‧‧‧ Entropy coding unit

107‧‧‧ Inverse quantization unit

108‧‧‧ inverse transform unit

109‧‧‧ Adder

110‧‧‧ post-processing unit

111‧‧‧Screen storage unit

112‧‧‧ In-frame prediction unit

113‧‧‧ Inter-frame prediction unit

200‧‧‧Image decoding equipment

201‧‧‧ Entropy coding unit

202‧‧‧Inverse Scanning Unit

203‧‧‧ Inverse quantization unit

204‧‧‧ inverse transformation unit

205‧‧‧ Adder

206‧‧‧ post-processing unit

207‧‧‧Screen storage unit

208‧‧‧ prediction unit in the frame

209‧‧‧ Inter-frame prediction unit

FIG. 1 is a block diagram of an image encoding device according to the present invention; FIG. 2 is a schematic diagram of an in-frame prediction mode according to the present invention; FIG. 3 is a schematic diagram of a scanning mode according to the present invention; and FIG. 4 is a diagram according to the present invention. A schematic diagram of the diagonal scanning of the invention; FIG. 5 is a schematic diagram of the scanning mode determined by the in-frame prediction mode and the size of the transform unit according to the invention; FIG. 6 is a block diagram of the image decoding device 200 according to the invention; And FIG. 7 is a flowchart of a method for generating a prediction block according to the present invention.

Claims (10)

A method for generating a quantized block includes: using a most probable mode (MPM) group to derive an intra-frame prediction mode of a current prediction unit, the MPM group including a left frame according to the current prediction unit The three intra-frame prediction modes determined by the intra prediction mode and the intra-frame prediction mode in the figure above; selecting an inverse scanning mode based on the in-frame prediction mode and the size of a transform unit; and inverse scanning according to the selected inverse scanning mode Significantly mark, coefficient symbols and coefficient levels to generate the quantized block, wherein the inverse scanning mode is one of diagonal scanning, vertical scanning and horizontal scanning, wherein when the prediction mode in the left frame and the upper frame are When the intra-prediction modes are the same as each other and the left-frame in-frame prediction mode is one of two non-directional in-frame prediction modes, the MPM group includes the two non-directional in-frame prediction modes and a vertical mode, and where , When only one of the in-frame prediction mode in the left image and the in-frame prediction mode in the upper image is available and it is one of a plurality of directional in-frame prediction modes, MPM group that contains the two non-directional picture frame prediction mode and a frame prediction mode available FIG. The method for generating a quantized block as described in claim 1, wherein if the size of the transform unit is equal to or smaller than 8 × 8, the inverse scanning mode is determined through the in-frame prediction mode. The method for generating a quantized block as described in claim 2, wherein horizontal scanning is applied to the vertical mode and a predetermined number of in-frame prediction modes adjacent to the vertical mode, and a horizontal mode and a mode adjacent to the horizontal mode are applied. A vertical scan is applied to the predetermined number of in-frame prediction modes, and a diagonal scan is applied to all other in-frame prediction modes. The method for generating a quantization block as described in claim 3, wherein if the size of the transform unit is 8 × 8, the predetermined number is 8. The method for generating a quantized block as described in claim 1, wherein if the size of the transform unit is larger than 8 × 8, a diagonal scan is applied to all in-frame prediction modes. The method for generating a quantized block as described in claim 1, wherein if the size of the transformation unit is 8 × 8, the significant mark, the coefficient symbol and the coefficient level are inversely scanned in units of a subset to generate a plurality of subsets, and A quantization block is generated by inversely scanning the subsets according to the inverse scanning mode determined by the in-frame prediction mode in the figure. The method of generating a quantized block as described in claim 6, wherein the number of the encoded subsets is determined using the position of the last non-zero coefficient and the selected inverse scanning mode. The method of generating a quantized block as described in claim 6, wherein a plurality of non-zero subset labels are used to determine the subsets to be generated, and the inverse scan mode is applied to the significant labels, coefficient symbols, and coefficient levels to generate the Subset. The method for generating a quantized block as described in claim 1, wherein when the prediction mode in the left frame is a DC mode and the prediction mode in the frame above is the DC mode, the MPM group includes the DC mode and a plane. Mode with this vertical mode. The method for generating a quantized block according to claim 1, wherein when the prediction mode in the left frame is a plane mode and the prediction mode in the frame above is the plane mode, the MPM group includes a DC mode, the plane Mode with this vertical mode.